Running DeepSeek with Grok search as an agent: 2.8M tokens, 94% cache hits, $0.04 per task
teortaxesTex · x · 2026-11-09
User teortaxesTex demoed a "DeepSeek harness with Grok search" setup: running DeepSeek as an agent that uses Grok's X search tool, with a screenshot of the full run.
- The trace shows 2.8M tokens consumed, 94% cache hit rate, $0.04 cost, 122 tok/s
- The model verified snowflake ID/timestamp consistency, traced quote relationships, and summarized discussion themes
- Caveats noted: Grok-only post access hits login walls, and the plugin itself is questionable (h/t @caliperdynamics)
A look at building cheap search agents with a non-frontier model plus tool access.
Related event: Developer Runs DeepSeek Agent with Grok-Powered X Search(2 posts)→
More from coding & agent
- RentAHuman lets AI agents hire real people for physical-world tasks via MCP API — conitzer · 2026-11-09
- Jane Street open-sources /socratic-tutor: an agent skill that teaches via questioning, not lecturing — daniel_mac8 · 2026-11-09
- Fly-by-Code: coding agents pilot drones via active visual and physical feedback loops — jbhuang0604 · 2026-11-09
- Unreal Engine 5.9 preview is the first UE release built primarily with agentic workflows — nicolascraske · 2026-11-09
- YC-backed Floworks launches Eniac, an API for verified large-scale web data extraction — ycombinator · 2026-11-09
- Mistral showcases Studio enterprise solutions from pentesting to KYC onboarding — MistralAI · 2026-11-09